How Fanatics moved from audience targeting to optimizing campaigns for customer LTV

However, after recent experiments, the sports betting operator found better results when it stopped defining its audience and adopted a more data-led approach, enabling it to increase projected customer lifetime value (LTV) by 19% from campaigns run with Cognitiv’s “full-funnel CTV solution.” 

According to statistics shared with Digiday, this activity around awareness of the sportsbook during marquee sporting events, such as March Madness, where the ad tech company’s solution drove an 11% lift in brand awareness and a 12% lift in customer LTV for Fanatics Casino.

Andy Magnes, director, paid marketing, Fanatics Betting & Gaming, told Digiday the uptick was the result of shifting from targeting predefined audience segments to employing tactics that optimize directly against customer value signals. 

The change reflects a broader evolution in digital advertising, where machine-learning systems are being tasked not merely with finding audiences, but with determining who those audiences should be in the first place.

Early work with Cognitiv focused on modeling audiences based on specific betting behaviors rather than optimizing campaigns for the lowest cost per acquisition.

According to Fanatics, these efforts improved on previous campaigns, with executives familiar with the developments telling Digiday the company ultimately achieved stronger performance when it stopped trying to define the audience itself and instead optimized for business outcomes.

Rather than relying primarily on demographic profiles or manually constructed audience segments, Cognitiv trains custom machine-learning models against advertiser-defined outcomes. 

“We just started modeling on value,” said Fanatics’ Magnes, adding that the more recent technique enabled it to better establish a link between ad spend and generating customer LTV. “You can certainly get ahead of yourself on targeting.”

The strategy was also aided by Cognitiv’s AudienceGPT product — a tool used prominently in the recent activity with Fanatics — which lets marketers describe prospective audiences in natural language, with the product then using machine learning systems to construct appropriate behavioral segments. These audience types can then be activated across different channels, such as online video or social platforms.

Jeremy Fain, CEO Cognitiv, explained to Digiday that advertisers are increasingly better served by defining desired business outcomes rather than applying hypotheses towards a predefined audience type.

“Fanatics tells us who the high-value people are, and we go out and find more of those people,” he said, describing a process whereby algorithms identify the characteristics associated with valuable customers instead of relying on marketer assumptions.

“Although at the top [of the marketing funnel], you do need a hypothesis [as to the target customer-base], and these prompts end up being that hypothesis,” added Fain.